Evaluating and changing representation in concept acquisition

  • F. Bergadano
  • F. Esposito
  • C. Rouveirol
  • S. Wrobel
Part 1: Constructive Induction And Multi-Strategy Approaches
Part of the Lecture Notes in Computer Science book series (LNCS, volume 482)


Adequate representation of examples and hypotheses is a key issue in concept learning. Simplistic representations may fail to allow for discriminant classification rules, while over-detailed inductive hypotheses may turn out to perform badly on new examples. If a representation is evaluated to fall in one of these two extremes, both causing poor performance, it is then necessary to change it. Change often depends on our knowledge of the predicates that are relevant in the application domain, but may also be automated in some cases. All of the above issues are analyzed in the present paper and methods for evaluating and changing a given representation are reviewed.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [Bis90]
    G. Bisson. K.b.g., a knowledge base generalizer. In Proc. of the seventh Int. Conf. on Machine Learning, pages 9–16, Austin, TX, 1990.Google Scholar
  2. [BS89]
    F. Bergadano and L. Saitta. On the error probability of boolean concept descriptions. In K. Morik, editor, EWSL89 — Proc. of the 4th Europ. Working Session on Learning, pages 25–36, Pitman, London, 1989.Google Scholar
  3. [Cha88]
    B. Chandrasekaran. From numbers to symbols to knowledge structures: Pattern Recognition and Artificial Intelligence Perspectives on the classification task. In E. S. Gelsema and L. N. Kanal, editors, Pattern Recognition in Practice, Vol. 2, North Holland, 1988.Google Scholar
  4. [DB89]
    L. DeRaedt and M. Bruynooghe. Towards friendly concept-learners. In Proc. of the 11th Int. Joint Conf. on Artif. Intelligence, pages 849–854, Morgan Kaufman, Los Altos, CA, 1989.Google Scholar
  5. [Esp90]
    F. Esposito. Automated Acquisition of Production Rules by Empirical Supervised Learning Methods. In M. Schader and W. Gaul, editors, Knowledge, Data and Computer Assisted Decisions, Springer-Verlag, 1990.Google Scholar
  6. [GS90]
    A. Giordana and A. Saitta. Abstraction: a General Framework for Learning. In Proc. AAAI Workshop on the Automatic Generation of Approximations and Abstractions, pages 245–256, Boston, MA, 1990.Google Scholar
  7. [MB88]
    S. Muggleton and W. Buntine. Machine invention of first order predicates by inverting resolution. In Proc. of the Fifth Int. Conf. on Machine Learning, pages 339–352, Ann Arbor, MI, 1988.Google Scholar
  8. [Mic83]
    R. S. Michalski. A theory and methodology of inductive learning. Artificial Intelligence, 20:111–161, 1983.CrossRefGoogle Scholar
  9. [Mit82]
    T. M. Mitchell. Generalization as search. Artificial Intelligence, 18:203–226, 1982.CrossRefGoogle Scholar
  10. [Mug88]
    S. Muggleton. A strategy for constructing new predicates in first order logic. In Proc. of the 3rd european working sessions on learning, pages 123–130, Glasgow, UK, 1988.Google Scholar
  11. [PS90]
    M. Pazzani and W. Sarrett. Average case analysis of conjunctive learning algorithms. In Proc. of the seventh Int. Conf. on Machine Learning, pages 339–347, Austin, TX, 1990.Google Scholar
  12. [Rou91]
    Celine Rouveirol. ITOU: Induction de Theories en Ordre 1 (Ph.D. Thesis. Technical Report, Lri, 1991. to appear.Google Scholar
  13. [Sch87]
    Jeffrey C. Schlimmer. Incremental adjustment of representations for learning. In Fourth International Workshop on Machine Learning, pages 79–90, Irvine, CA, June 1987.Google Scholar
  14. [SSS88]
    W. Siedlecki, K. Siedlecka, and J. Sklansky. Mapping techniques for exploratory pattern analysis. In E. S. Gelsema and L. N. Kanal, editors, Pattern Recognition and Artificial Intelligence, North Holland, 1988.Google Scholar
  15. [Sub90]
    Devika Subramanian. A theory of justified reformulations. In D. Paul Benjamin, editor, Change of Representation and Inductive Bias, pages 147–167, Kluwer, Boston, 1990.Google Scholar
  16. [Utg86]
    Paul E. Utgoff. Shift of bias for inductive concept learning. In R.S. Michalski, J.G. Carbonell, and T.M. Mitchell, editors, Machine Learning — An Artificial Intelligence Approach, pages 107–148, Morgan Kaufman, Los Altos, CA, 1986.Google Scholar
  17. [Vap74]
    V. Vapnik. Estimation of dependencies based on empirical data. Springer Verlag, New York, 1974.Google Scholar
  18. [Wro88]
    S. Wrobel. Automatic representation adjustment in an observational discovery system. In D. Sleeman, editor, Proc. of the 3rd Europ. Working Session on Learning, pages 253–262, Pitman, London, 1988.Google Scholar
  19. [Wro90]
    Stefan Wrobel. Concepts and Concept Formation: Fundamental Issues. Arbeitspapiere der GMD Subreihe KI, GMD (German Natl. Research Center for Computer Science), P.O.Box 1240, 5205 St. Augustin 1, FR Germany, 1990. to appear.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1991

Authors and Affiliations

  • F. Bergadano
    • 1
  • F. Esposito
    • 2
  • C. Rouveirol
    • 3
  • S. Wrobel
    • 4
  1. 1.University of TorinoItaly
  2. 2.University of BariItaly
  3. 3.LRIUniversité de Paris SudFrance
  4. 4.GMDBonn

Personalised recommendations